The Determinants of Electoral Reform: A Synthesis of Alternative Approaches

Transcription

1 DRAFT March 2009 Please Do NOT Cite Without Authors Permission The Determinants of Electoral Reform: A Synthesis of Alternative Approaches Gideon Rahat & Reuven Y. Hazan Department of Political Science The Hebrew University of Jerusalem Prepared for delivery at the workshop on Why Electoral Reform? The Determinants, Policy and Politics of Changing Electoral Systems European Consortium for Political Research s Joint Sessions of Workshops April 2009, University of Lisbon, Portugal

2 1 The Determinants of Electoral Reform: A Synthesis of Alternative Approaches Introduction If one still needs to be convinced that no single approach succeeds in explaining the determinants of electoral reform, one should read three recently published books (Blais, 2008a; Rahat, 2008; Renwick, 2009). If, however, one is more easily convinced, then the eight page analysis by Katz (2007) should suffice. This paper adopts a similar conviction, and attempts to find a way around the problem by synthesizing between several approaches within a comprehensive framework for analysis. This framework proposes a menu of barriers which those who aspire to reform the electoral system are likely to face, and will need to pass in order to achieve their goal. Katz (2005) picked an inherently contradictory title for his chapter on electoral reform, Why are There so Many (or so Few) Electoral Reforms? This title expresses the two sets of opposing logical expectations that one encounters when studying electoral reform. On the one hand, we should expect electoral reform not to occur at all, because the ones who hold the key for change are those who won under the rules of the exiting electoral system, and thus they should have no interest in changing the rules of a game in which they are successful. For example, the two biggest parties in a majoritarian electoral system should avoid any calls for adopting more proportional elements, and the smaller parties in a PR system should block any attempts to implement aspects of majoritarianism. On the other hand, we should

3 2 expect reform to occur frequently, because those that hold the key for change should reform the system in a manner that will be to their benefit. For example, we might expect a party in a majoritarian system to adopt PR if it expects electoral defeat, in order to minimize damages. Or, we might expect the two largest parties in a multiparty system under PR to collude and create a duopoly by adopting a majoritarian electoral system. Renwick (2009) shows that indeed we have two kinds of electoral reform. The first kind, the elite majority imposition is advanced by the elite in order to manipulate future electoral results (France 1951, 1984, 1986 and Italy 1953, 2005), which leads us to the question: Why so little? And the second kind, the elite-mass interaction, is the one in which reform is adopted against the will of the elite (Italy 1993, New Zealand 1993, Japan 1992 and arguably Israel 1992), which leads to the question: How come? Linking these notions to Rahat s (2008) analysis of the arena in which electoral reform is promoted and (sometimes) adopted, we can talk about two settings (Figure 1). The first (the left hand side) is one in which power is concentrated in the hands of a single actor, who can choose to reform the system on his own. The second (the right hand side) is one in which power is dispersed among many actors, and no one actor can single handedly block a reform. In between those cases we have normal democratic politics, in which reform is blocked because of what we usually call checks and balances. In terms of veto actors, reform is likely to occur either when there is only a single veto actor or when there are so many potential veto actors that one of them, on its own, cannot veto reform.

4 3 Figure 1: Reform Occurrence in Substantially Different Settings We suggest that scholars should analyze reform using a barriers approach. This approach first asks what keeps the electoral system stable over time. It lists seven possible barriers that reformers may face when trying to promote electoral reform. This menu of barriers is based on two families of approaches, the institutional approach and the rational choice approach. Each barrier is presented separately, and empirical evidence is provided to examine its significance. After all, it is likely that at any given moment there is someone who is interested in electoral reform. The problem is thatt of promoting reform and turning it from an idea into a reality. In order to reform an electoral system, reformers need to recruit the support of a majority, and many times of majorities. We ask what stands between them and the recruitment of majority support for electoral reform? We subsequently examine the ways reformers passs each barrier. They use an array of tactics and strategies that challenge the status quo. These include attempts at changing the playing field, proposing multifaceted mixed systems and selecting the right timing. The barriers approach synthesizes several approaches and is useful because it can be used to explain both the successs as well as the failure of reform initiatives, and even those casess of non-reform.

5 4 The Barriers Approach Table 1 presents a list of barriers that reformers are likely to face in their attempt to promote their cause. In this section, each barrier is described and then checked by looking at cases that point out both its significance and its limitations. Table 1: Barriers to Electoral Reform in Democracies The Barrier Procedural superiority of the institutional status quo Political tradition Societal structure Systemic rationale Actors vested interests Coalition politics A majority for change; disagreement over content The Approach Old institutionalism Cultural institutionalism Sociological institutionalism New institutionalism Rational choice (group/individual level) Rational choice (game theory) Rational choice (game theory) Source: Elaborated from Rahat (2008). Institutional Approaches Institutions are largely stable and are thus useful anchors for researchers attempting to analyze complex realities. This allows institutionalists to treat institutions, most of the time, as independent variables, which explain rather than have to be explained. In our case an institution, the electoral system, is treated as the phenomenon that has to be explained the dependent variable. Some of the institutional approaches outlined below look at other institutions (reform procedures) for the explanation (old institutionalism); others look at the interaction between culture (cultural institutionalism), society (sociological institutionalism) and political institutions; while still others (new institutionalism) look at other institutions (party systems) or even the functioning of electoral systems themselves as an explanation for their stability or change. Before turning to deal with each approach, it should be noted that none of the scholars who proposed these approaches, or used

6 5 them, suggest that they can fully predict where and when reform will be adopted, or its specific characteristics. Old Institutionalism Old institutionalism (what we might also call today Constitutional Law) focuses our attention on the procedural path that reform faces before it is adopted. Since we are dealing with electoral reforms within democracies in which the rule of law is respected by definition this element should be treated seriously. In this context, the institutional status quo is the default. The reformers, by contrast, face several procedural barriers in their attempt to change the institutional setting, and if they fail to pass each and every one of them, they return to the starting point. In other words, the institutional status quo has inherent primacy over reform, and the amendment procedures supply the guardians of the status quo with numerous opportunities to block reform. Reformers cannot allow themselves to be defeated in even one of these battles. Electoral reform should thus occur only as a result of an intended and determined political action by actors ready to face these difficulties. Even in those cases where electoral reform can be legislated as a regular law, reform is not about a single successful attempt to recruit a majority. Rather it is about maintaining and rebuilding majorities at different times (i.e. several parliamentary readings) and within different forums (i.e., plenum, committee) whose composition is not necessarily similar. Moreover, the guardians of the status quo can attempt to block reform not only by simply recruiting majority support for the status quo, but also by

7 6 taking advantage of opportunities to delay the progress of reform or by dividing the reform camp. 1 In many countries, however, electoral reform requires more than regular legislation, such as an amendment to the constitution or an amendment to special laws. 2 These may involve recruiting particular majorities (from an absolute majority up to twothirds and more) or securing the consent of concurrent majorities. In those cases, reform becomes an even more demanding task. Based on the above discussion, we hypothesize that: The more demanding the electoral reform process, the less likely that electoral reform will be adopted. The electorate, through a referendum, can block reform, as was the case in Ireland in both 1959 and A party with a majority in parliament, Fianna Fail, legislated a law that replaced the PR system with a plurality system, but because PR appears in the constitution, they still had to face a referendum and there the majority of voters rejected the reform (Gallagher, 1996). In other cases (Italy 1953, 2005, France 1951, 1986, 1988), where reform needed only the support of a majority in parliament, the elite-imposed reform was adopted. The size of the majority needed can also be crucial. In British Columbia, a reform that was recommended by a citizen assembly won the support of the majority of the voters, 57%, yet was rejected because it was 1 2 Postponement of a reform bill, for example, by the chair of a parliamentary committee or by the speaker, can help reform opponents to delay a decision until the timing is better for them, i.e., as public pressure for reform decreases. A long enough delay may lead to the slow death of reform as other issues capture the agenda. Another strategy is that of proposing an alternative bill, or introducing an amendment that divides the supporting majority for example, the time-frame for implementing the reform. Special laws have a higher status. Their amendment requires a larger majority than regular law, but lower than the constitution.

8 7 previously determined that a majority of 60% would be needed (Carty, Blais and Fournier, 2008). In Italy, the abolition of the PR element in the mixed-member majoritarian system won the support of 91.5% of the voters in one referendum (1999) and 82% in the next (2000), but because turnout was lower than the needed 50% 49.6% in 1999 and 32.4% in 2000 the initiative was rejected (Uleri 2002). New Zealand, on the other hand, adopted reform with a majority of only 53%. A comparative study indeed demonstrates that there is a significant correlation between the level of difficulty set by the constitutional amendment formula and the frequency of the adoption of constitutional amendments (Lutz, 1994). Yet, this study, as well as other comparisons (Duchacek, 1973), demonstrate that procedures are far from an explanation of all the variance, and that other factors should be used to explain differences in constitutional stability. It should also be noted that in some of the cases where referendums or initiatives were used (i.e. Italy 1991, 1993 and New Zealand 1992, 1993), in addition and prior to the parliament decision, they were conductive for reform because once the people gave their verdict, the parliament found it hard to ignore it. Even if procedures cannot tell the whole story, they should be analyzed because they determine the path or (possible) paths for reform, the magnitude and kinds of majorities that have to be recruited, and even, in some cases, the opportunities for reformers. As the comparison between the British Columbia and the New Zealand cases demonstrates, the size of the needed majority may even account for the success or failure of an electoral reform.

9 8 Cultural Institutionalism According to this perspective, the institutional setting especially that of stable, established democracies is linked to a country s culture and its political tradition. It is culture and political tradition which explain the tendency of the Anglo-American democracies toward majoritarianism and the tendency of Continental democracies toward consensualism including their PR electoral systems (Lijphart, 1994; 1998). In this perspective we expect reform or even a serious consideration of it to appear as a result of fundamental political and social change. At all other times it is difficult to mobilize majorities behind support for the reform of a system that is seen as linked to the political tradition. Such traditional values include the inclusiveness and representativeness of PR, as part of the political culture in most Western European consensus democracies, or the constituency representation of the Anglo-Saxon democracies. Like other cultural elements, these might be built on myths rather than facts (e.g. constituency representation in the US compared to that in the UK), but they still have a power of their own. We thus hypothesise that: The older an electoral system, the less likely that electoral reform will be adopted. Indeed, the large wave of reform and the serious deliberations surrounding it even in some countries that kept their electoral systems occurred at the time that countries broke with the past and adopted universal suffrage (Carstairs, 1980). PR replaced majoritarian systems as part of a wider transformation, and was seen as a part of the democratization process, a move that ensured the principle of one man, one vote (Blais, Dobrzynska and Indridason, 2004).

10 9 Since the Second World War we have witnessed only a few electoral reforms in the established democracies. Countries that adopted electoral reform include France, which is not necessarily an example of stability, but rather a country in which two traditions (although in changing forms) continued to struggle and shake the foundations of the republic. The examples also include the relatively young democracies of Italy, Japan and Israel. One may argue that the young age of democracy might make it more prone to reform. After all, not all relatively young established democracies have reformed their electoral systems. And, one of the most veteran democracies in the world (maybe the oldest, if we use universal suffrage to signify the starting point), New Zealand, also reformed its system. Cultural institutionalism is thus a passable barrier that reformers should take into account when proposing and promoting a reform. Sociological Institutionalism This perspective looks at the relationship between the institutional setting and the structure of society. The stability of an established democracy is thought to be the result of the wise adoption of an institutional setting that fits the structure of society. In a divided and heterogeneous society, the expectation is that power will be shared through the adoption of a consensus regime, and a central feature of such a regime is a PR electoral system. This arrangement allows various social groups to live within a common democratic framework. By contrast, a homogenous society can allow itself to adopt a majoritarian structure in which majoritarian electoral systems are a main feature because it can avoid slicing the pie into many pieces (Lijphart, 1984).

11 10 According to this perspective, we expect a country to preserve its electoral system as long as it fits the structure of its society. We hypothesize that: The more a country s electoral system fits the structure of its society, the less likely it is to adopt electoral reform. Indeed, Lijphart (1987) thought that New Zealand, which was located on the majoritarian pole, should adopt a more consensusual regime. If we follow the logic of Lijphart s modeling, New Zealand was the natural candidate for change, as the other SMP (single member plurality) democracies (US, Canada, Australia) have consensus features that compensate, like federalism. Moreover, Jackson (1994) in line with Lijphart s theory claims that the evolutionary transformation of New Zealand s society from homogeneity into a multicultural society explains the pressure to reform the majoritarian model. Lijphart (1993) also claimed that following his analysis, Israel should reform and adopt a somewhat less proportional system. Indeed Israel reformed its system, although unfortunately not the way recommended by Lijphart (Hazan 1996; 1997). While not necessarily agreeing with Lijphart s analysis, Medding (1999), Sheffer (1999a/b) and Hazan (1998) all pointed out that a change in power distribution among political and social forces preceded reform in Israel. Reforms can thus be interpreted as an attempt either to adapt to societal changes (New Zealand) or to counter their influence (Israel). Indeed, pressures for reform increased both in Italy and in Japan when the old justifications for preserving a malfunctioning and corrupt system were removed. The decline of the significance of communism (and anti-communism) alongside religion in Italy, especially after the collapse of the

12 11 Soviet bloc, are factors that led to the re-examination of the Italian institutional order. The end of the economic miracle in Japan also lead to a re-examination of its institutional order. This approach also has its limits. It does not tell us when such a reform will occur, who will carry it out, and why countries with seemingly the same societal structure do not choose the same institutional structure. We thus expect reformers to address sociological institutionalism when proposing a reform, and to have answers for those who will criticize their initiative from this perspective. New Institutionalism System Level Rationality According to this perspective, a country should preserve its electoral system as long as it produces certain expected outputs. Unlike the previous approach, it does not address the social setting but focuses only on the institutional setting, namely the party system and the electoral system itself. Colomer (2005) claims that electoral system stability, or change, results from the number of parties: in democracies with a low number of effective parties among voters a majority system will be adopted and preserved, while in democracies with a higher number of effective parties among voters PR will be adopted and preserved. If we deduce a barrier here, we hypothesize that: The more an electoral system fits the number of effective parties among the voters, the less likely is it to be reformed.

13 12 Blais, Dobrzynska and Indridason (2004) claim, in line with Colomer, that the countries with multiparty systems and coalition governments as expected from PR were those that actually adopted PR at the beginning of the 20 th century. Renwick (2009), on the other hand, does not find the linkage that Colomer found in his analyses of the post 1945 cases. Shugart (2001; 2008) proposes two somewhat different scenarios for the creation of systematic pressures for electoral reform, from which it is possible to deduce rules for the stability of an electoral system. According to Shugart (2001), we expect to see stability in countries that have an electorally efficient system. An efficient system is defined as a system that succeeds in translating the will of the majority of the voters into policies. The efficient systems are those that balance the positive properties of both PR (representation) and majoritarianism (clear choices), alongside those of both personal (accountability) and party-centered (a comprehensive program) electoral systems. From a systemic perspective, it is rational to preserve such a balanced existing order. Shugart (2001) claims that non-efficient electoral systems (he called these electoral systems extreme ) are more likely to face reform pressures, but reform will not necessarily occur under such circumstances. The hypothesis is thus: The more efficient a country s electoral system, the less likely it is to adopt electoral reform. In all four cases of electoral reform in the 1990s, Shugart (2001) identifies an extreme feature that the reforms tried to correct. New Zealand s extremist pluralitarian electoral system allowed parties whose support declined to implement radically

14 13 contested policies (Vowels, 2008). That is, the pathology of elected dictatorship was strongly pronounced during the push for reform. In Israel, a hyper-representative electoral system produced coalition politics that led to difficulties in electing and identifying a responsible government, especially after the stable and effective dominant party system was replaced by a highly competitive and seemingly unstable multiparty system (Rahat, 2001). In Japan, a candidate-centered electoral system was replaced with the hope of reducing the incentives for personally corrupted politics (Reed and Thies, 2001). In Italy, a candidate-centered, hyper-representative electoral system was replaced with a mixed-member majoritarian system. The aim here was to decrease corruption, clientilism and malfunctioning by creating a two-bloc system with reduced incentives for personal corruption (Katz, 2001). Shugart makes the case that this approach is limited in that it can predict pressures for reform but not the actual implementation of reform some countries that had extreme systems implemented reform while others did not. Indeed, Several detailed historical accounts of the failed attempts at electoral reform in the UK (Farrell, 2001), Canada (Weaver, 2001) and the Netherlands (Van der Kolk, 2007) demonstrate that even if there are strong claims concerning the faults of the system (and even if there are opportunities for reform) the status quo is still likely to prevail. In a more recent study, Shugart (2008) did not relate pressures for electoral reform to the distance between an electoral system and a universal type of balanced, efficient electoral system. Instead, he suggested that reform is more likely (to be seriously discussed and even, rarely, adopted) where the electoral system does not adhere to its own promises. That is, an SMP electoral system is likely to create reform pressures when it manufactures a majority for the party with the second highest number of

15 14 votes, rather than for the first; or when it fails to create a significant opposition by harshly under-representing the second largest party. From this, we can deduce the claim that a system that adheres to its promises is likely to be preserved. The hypothesis is thus: The more that a country has an electoral system that fulfills its implicit expectations, the less likely is it to adopt electoral reform. The experience until now, as Shugart (2008) himself demonstrates, is that the failure of SMP to meet expectations created several windows of opportunity to promote and seriously consider reform. In only one case so far, did it lead to reform (New Zealand). This, in itself, does not invalidate the proposition above, as consideration of reform is of course a necessary condition for its adoption (though not a sufficient one). What is still missing is an application of this approach to other majoritarian electoral systems (two round, alternative vote) and to the whole family of PR electoral systems (Shugart, 2008: 56 n3). Moreover, this approach seems to contradict Shugart s earlier outlook (2001), which claimed that the pathologies inherent to the system are likely to lead to its reform, rather than its failure to meet expectations. Rational Choice Approach(es) The rational choice approach sees the stability of electoral systems as a result of the calculated behavior of politicians. The preference for the status quo is the result of a comparison between the existing system and alternative systems. Indeed, a study of attitudes of politicians in four established democracies towards reforms (term limits, the use of referendums and citizen initiatives, and compulsory voting) showed that the politicians' attitudes toward these issues are affected by their positions those who

16 15 won the elections and who are in a position to implement reform showed less support for electoral reforms than the losers (Bowler, Donovan and Karp, 2006). Pilet (2008b) studied the attitudes of 59 parties in eleven attempts of electoral reform and found that retrospective (how well they did, and do, under the existing rules) and prospective (how well they are expected to do under alternative proposed rules) motivations taken together help to explain the position of political parties towards reform. Rational Choice: Vested Interests According to Benoit (2004: 379), Electoral systems should cease to change once no party or group of parties with the fiat power to change the electoral law perceives a potential seat gain by doing so. Parties who are favored by the electoral system, in its translation of votes into seats, are likely to have strong incentives for supporting the existing system. When we talk about vested interests in electoral systems, we should first delineate between kinds of systems, because different parties (distinguished mainly by their size) are rewarded or punished by different systems. We will, therefore, discriminate between majoritarian and PR systems. We will also distinguish, when necessary, between SMP (single member plurality) and other majoritarian systems. As Blais, Dobrzynska and Indridason (2004) demonstrate, in non-smp systems it is hard for parties to identify their specific interests, while in an SMP system these are rather clear. SMP systems supply valuable rewards in terms of representation and influence to its successful players (Blais, 2008b). The party that holds a majority of seats enjoys a

17 16 large bonus, usually a manufactured majority. That is, the majority that is supposed to reform the system has the strongest interest in its preservation. The second largest party benefits from the fact that the system encourages a duopoly, and hopes to enjoy, sooner or later, the rewards of the front-runner. Thus, even the less successful party tends to stick with the system (Blais and Shugart, 2008). We also expect smaller parties with a regional orientation to prefer SMP, since they are favored due to their vote concentration. Blais and Shugart (2008) claim that SMP is still with us as a result of its large rewards, because in the normative battlefield, when it is pitted against the alternatives in citizen assemblies or in expert committees moderate PR systems usually have the upper hand. Thus, we hypothesize that: SMP systems will be preserved since they serve the interests of those who hold the key to their reform. Boix (1999) examined the wave of reform in majoritarian systems at the beginning of the 20 th century. He explained them as a result of the strategy of the ruling parties to change the system in an effort to control its results adopting PR in order to control possible negative results, namely the creation of a manufactured majority by the growing Socialist parties due to the extension of universal suffrage. In the post 1945 cases, electoral reform penetrated the political agenda of countries with majoritarian systems only when one of the two large parties felt that it was continuously prejudiced by the majoritarian system (Shugart, 2008). Since the Second World War including the New Zealand case there were no instance of major parties in SMP systems acting to promote reform. Yet, two elite

18 17 imposed reforms of a majoritarian system did occur. One was the replacement of the two round system in France with PR in 1986, mainly aimed to reduce the expected electoral damages by the ruling Socialists. The second was the replacement of the MMM Italian system by PR with a bonus for the largest party alliance. Thus, there are instances in which parties that benefitted from the biases of majoritarianism expect their rivals to win and gain from them, and thus in order to minimize this they reform the electoral system. Can we argue that SMP systems (and also majoritarian systems in general) are prone to change when those who benefit from their biases predict that their rivals are going to gain the upper hand? We can claim that, except in the New Zealand case, majoritarian systems were reformed by parties aiming at controlling a predicted loss of votes. Yet, we must remember that in most cases where parties predicted a loss under majoritarian systems, they did not reform the system. It may be because of legitimacy constraints a party fears that playing with the rules of the game will lead to a loss of legitimacy that in turn will lead to further electoral loss (Renwick 2009). Or, possibly, losing parties in majoritarian systems have a good reason to stick with a system that preserves the duopoly and also promises them that one day they will be back as the governing party, without having to share power with coalition partners. What about PR systems? Following the seat maximization logic, we should expect PR to survive due to the existence of seat-maximizing partisan vested interests under two scenarios. First, when there is a dominant party system, but the large party does not possess a majority of the seats. In this case, the smaller parties that together have the

19 18 majority prefer the way that PR translates votes into seats over a manufactured majority for the large party. Thus, we hypothesize that: A reform in a country with a PR electoral system is not likely to occur as long as it has a dominant party which does not have a majority of the seats. The fact that democracies that had such a party system did not reform does not really say much, since reform is in any case a rare phenomenon. However, if a dominant party in such a scenario did attempt to promote reform, but was blocked by all the other parties, then there is an interesting case here. Indeed, that is what happened in Israel in the 1950s. The leader of the dominant Mapai party called for the adoption of an SMP system. Almost all the smaller parties not only blocked this attempt, but gave the electoral system a constitutional status that no other law in Israel enjoyed at the time protection from an amendment by an absolute majority (Rahat, 2008). The second scenario in which PR is likely to be preserved is a multiparty system with a high effective number of parties, and in which there are no two parties that are relatively large and together posses the needed majority for reform. If there were two such parties, they are expected to adopt majoritarianism with its incentives for strategic voting. Similar to the scenario above, the smaller parties that together have a majority in the parliament have a vested interest in the existing system. While the expectations above make sense, we still need to tackle two additional scenarios in which one or two parties have the needed majority for reform. What

20 19 happens when one dominant party in a PR system has a majority of seats and holds the key for a change that will likely enhance its majority? What happens when two relatively large parties posses a majority of seats and hold the key for an electoral reform that will likely enhance their status, and may even create a duopoly? We expect a reform in a country with a PR electoral system under either scenario a dominant/single party, or two relatively large parties, hold the majority needed for reform. This expectation is, of course, absurd if we assume that it can predict reform universally. Yet, it is an excellent tool in order to assess why parties preserve certain electoral systems (or fail to reform them) in which they do not have vested interests, at least in terms of seat maximization. We can make a strong claim, within the realm of the rational choice approach, that stability can be explained as a result of the tendencies of politicians to be risk averse. If they predict that a reform will serve them, the politicians have to estimate what their future support will be and what will be the support of their rivals. In other words, there is always uncertainty concerning the uncontrolled behavior of voters and other parties (and possibly additional actors). In his analysis of the behavior of the two largest Belgian parties, Pilet (2008a: 42) shows that these parties would prefer to maintain a safe status quo they are satisfied with than to speculate in order to accumulate. He thus suggests that we assess whether parties are satisfied with a system before evaluating their being the potential winners or losers of an electoral reform. As noted earlier, we also have empirical evidence that those in power have stronger preference for the institutional status quo (Bowler, Donovan and Karp, 2006). Stability of a system may also result from the politicians' fear that attempts to

21 20 maximize their power through changes in the electoral rules will be counterproductive, because the public will react negatively to what looks like an attempt to manipulate the rules of the game (Katz, 2007). In addition, it is not only parties that have vested interests in maintaining the institutional status quo, but also individual incumbents and strong interest groups that are likely to guard the system from change (Blau, 2008; Dunleavy and Margetts, 1995). However, we have cases in which a party (LDP in Japan) or two parties (Likud and Labor in Israel) that both had the majority needed for reform and actively promoted it subsequently failed to implement it. Why did they fail? This leads to the next possible (additional) barrier, namely coalition politics in their wider sense, not only between parties but also within party alliances and even inside parties (among factions). Rational Choice Coalition Politics and the Veto of the Few In PR systems, large parties have most of the time good reason to believe that they will gain both representation and influence from the adoption of majoritarian features. What blocks these large parties from reforming the system is an immediate interest in maintaining their coalitions. In simple terms of seat maximization, it cannot be understood why two parties that have a majority of the seats in a PR setting, and can thus reform the system to their benefit, refrain from doing so. The answer to this puzzle is to be found within the realm of rational choice if we are ready to go beyond seat maximization. We need to see parties as entities that are interested more in governing than in possible future seat maximization. For example, in Israel where the two large party alliances held the

22 21 needed majority there was stable majority support for electoral reform both among legislators and the public for decades. Yet reform was never adopted because of coalition pressures (Rahat, 2008). Diskin and Diskin (1995) analyzed the Israeli case and showed that this veto power was not limited to parties in minimal winning coalitions. Electoral reform progressed further than ever before in the grand coalitions of , which included both large parties and other smaller ones, but reform was nevertheless blocked. While a small party could not bring down these coalitions, it could nevertheless influence the behavior of its larger partners by offering its support for a narrow coalition led by one of the large parties who would reject reform, or by threatening to refrain from supporting it in future coalitions. A similar explanation is given also in the West German case of reform promotion by the grand coalition of the 1960s (Renwick, 2009). Based on Van der Kolk s (2007) analysis, it seems that coalition politics also played a role in preserving the stability of the highly proportional electoral system in the Netherlands. Such a scenario occurs not only in the context of inter-party politics. A single party within an alliance of parties, or even an intraparty faction, can play the role of a veto actor. This happened when the Japanese LDP, which held a majority in parliament, failed to promote a reform in 1970s due to the opposition from some of its factions (Renwick, 2008). It was also the case in Israel in the period, when parties within alliances blocked reform promotion (Rahat, 2008). Coalition politics make reform difficult, in that any crucial member (an intraparty faction or a coalition partner) of the coalition (governing coalition, or even a policy coalition on an important issue) holds veto power against reform. In other words, even

23 22 if reform has majority support in the legislature, it is traded (willingly or under pressure) by part of this majority in order to maintain the coalition government (or to sustain party cohesion). Rational Choice (Game Theory): Majority for Change, Disagreement on Content Another possible barrier for reform is the need to reach an agreement over the content of the reform among the needed majority. Here we might see a struggle within the reform camp, in which several actors try to promote their specific version of reform. As long as these forces do not agree among themselves, the status quo, even if supported by a mere minority, will prevail. Lack of an agreement can lead to the postponement and burial of reform bills. Without an agreement on the substance, the institutional status quo could be a Condorcet winner that is, an option that is supported by a majority in all pair-wise matches. This will happen when the alternatives are rejected, one by one, by the temporary coalitions of status quo loyalists (by themselves, only a minority) and those supporting a different version of reform (Browne and Hamm, 1996). While a majority tends to prefer various reform initiatives to the status quo, each camp refrains from supporting the reform versions of the other camps (although preferable to the status quo) because it hopes that the others will yield and support its specific version of reform, thus leading to the failure of reform. The Relative Strength of the Barriers and the Ways to Overcome Them According to the approach proposed here, reforming the electoral system requires passing all of the barriers mentioned above. That is, a reform initiative will be blocked if it cannot cope with each and every barrier. In this section, we look at the

24 23 tactics and strategies used by reformers to pass these barriers. Reformers must be determined in the sense that reform will not just occur, nor will it face difficulties similar to most other legislation. After all it is about the political fate of those who legislate it, and it is not just policies. Once they are present, these determined reformers face the barriers we outlined above. We start with those barriers that were found by Rahat (2008) to be the less challenging (tradition, social structure, system-level rationale). Then we look at those that were found to be effective in delaying reform (procedures and the need for agreement over content). These barriers required the reformers to invest much effort in keeping reform on the agenda and in promoting it, and to be prepared to make the necessary compromises in order to sustain and rebuild their coalitions. Otherwise, accumulated delays will often result in a reform's burial. Finally, we look at the main barriers, the vested interests in the electoral system which is especially relevant for majoritarian system, especially SMP and that of coalition politics which is relevant for most cases in a PR setting. One should note that for the more complex cases of electoral reform, those that were not single handedly promoted by the elite (e.g. the electoral reforms of the 1990s), the rational-choice derived barriers (vested interests, coalition politics, disagreement on reform content) are much more central then the majority imposed reforms. The various institutional barriers, however, can be equally relevant in these two kinds of scenarios.

25 24 Those Barriers which Political Science Thinks are Important A) Culture, Societal Structure and Systemic Rationale Ostensibly, these barriers are somewhat more difficult to pass when the existing electoral setting is PR and not majoritarian (Blais and Shugart, 2008). However, the electoral reforms in the 1990s (Italy, Japan, New Zealand, Israel) showed these barriers to be quite weak for two reasons. First, the feeling was that of crisis, and in such times, when the system is seen as dysfunctional, it was easier to promote change. In such a situation, the opponents of reform were seen as motivated by their selfcentered interests, while the reformers championed the systemic values. Moreover, all the proposals were for mixed systems (Shugart and Wattenberg, 2001) which combined both the old and the new, demonstrating sensitivity to those elements that were perceived as positive in each country s political tradition, and thus they claimed to be taking into consideration the societal structure and the systemic virtues. To different audiences, and at different times, reform could be marketed as preserving or changing, fulfilling one value or another or any of these simultaneously. To put it bluntly, the easier barriers to overcome were those that scholars usually perceive as the most important for the well-being of the polity: the suitability of the electoral system to the country s political culture and to the structure of its society, or the best possible electoral system from a systemic perspective. In those cases where reform was imposed by the elites, these barriers were somewhat more difficult to pass. Elite-majority (Renwick 2009) imposed reforms were justifiably perceived as being a self-serving manipulation of the system. This gave power to of the opponents who used tradition/culture, societal structure or systemic arguments. Moreover, the promoted reforms were usually not of the mixed kind that

26 25 could have been a good platform to answer such criticism. Thus, passing these barriers required the elite to pay a price in order to adopt reform, contrary to the case of elite-mass interaction where the price was paid by those not willing to promote reform. It is thus not a coincidence that elite-imposed reforms passed in countries with relatively shaky politics like Italy and France, where manipulation of the rules are much more tolerated. Yet, even in these cases, some initiatives were put aside, or abolished, after the elite felt that it would pay a price in terms of public support. In Germany, this legitimacy constraint was central in explaining the stability of the electoral system in which the two large parties could unilaterally adopt majoritarian features (Renwick 2009). 3 B) A Delaying Barrier #1: Procedures There are several possible strategies to cope with the procedural barrier. One is to choose, if there are several procedural paths for reform, the one that is likely to supply reform with its best chances. In Italy, for example, the initiative path bypassed the possible rejection of the politicians. Another way is to create an alternative path. In New Zealand an ad hoc referendum enabled reformers to bypass the strong vested interests of the politicians in the SMP system. In Israel, the very threat of nominating an external committee, or conducting a referendum, helped to pass reform through the first reading. Finally, reformers can widen the arena for reform (as they did in the 1990s), involving non-formal actors and creating pressure on insiders. This means that the field in which the reform game is played is wider than the one formally defined by law, and thus there is additional space to maneuver (Rahat 2008). 3 We refrained from relating to Colomer s (2005) claims here because his approach is deterministic and does not leave room for intended human action. Yet, the gap between voting and seat allocation can give reformers, or opponents of reform, some ammunition in their struggle.

27 26 C) A Delaying Barrier #2: The Need for Agreement on Reform The cure for this problem is naturally a compromise. Such a compromise means that the electoral system should be seen as serving the interests of various potential reform supporters. We are likely to end up, once again, with a mixed system because its complexity and multifaceted nature enables the reformers to promise many things to many forces. Another way in which reformers may try to cope with the need to compromise is to start with promoting ambivalent bills, with the hope that later on all reform supporters will feel obliged to find a compromise since they are obligated to the cause of reform. This tactic, however, does not work many times, because politicians usually demand guarantees before they commit (Rahat 2008). D) A Main Barrier #1: Vested Interests Vested interests are a powerful barrier. Asking politicians to reform may be seen as asking them to put themselves in jeopardy. One way to cope with vested interests is to change the power balance, to recruit actors that do not have vested interests in the system but have an interest in the overall functioning of the system. Interest groups, public opinion and experts can be recruited from outside to shake the confidence of those who hold the key for change. Renwick (2009: 25) captures this path elegantly as a tri-stage process: First, a minority of politicians, for various reasons, come to see reform as desirable. Lacking the power to enact change through regular legislative channels, these pro-reform politicians, at the second stage, take the issue to the public. Attempts to spark public interest in the esoteric subject of electoral reform are unlikely in general to succeed. But where the public is angry at the state of the political system and electoral reform can plausibility be portrayed as a (partial) solution to the problems prompting that anger, public pressure for change in the electoral system can be stocked. At the third stage, this pressure forces the reluctant majority among politicians to accept reforms they do not want.

28 27 Both the limit and the potential concerning the rational choice model in general, and the seat maximizing interpretation in particular, are well accounted for. 4 We suggest that the problem of incomplete information must be taken into account here. Vested interests are stronger barrier when the parties have information on the working of the system. Rahat (2006) compared the motives for reform adoption and abolition in Israel, and found that that most of those who supported reform ended up suffering from it while those who rejected it actually benefitted from it. After reform was adopted, and its consequences became evident, those that gained from the reformed system opposed its repeal and those that suffered from its consequences supported its abolition. The standard rational choice assumption, that politicians are indeed seat maximizers, became valid. Calculations are dependent upon information; when information is missing or manipulated, politicians might vote against their objective interests. In her study of the Latin American cases, Remmer (2008) found evidence that electoral changes were conducted in a strategic manner, in a way we would expect them to be following what we know about the influence of electoral systems. Electoral reform can be promoted if political actors can be convinced that other interests might be well served through their support for electoral reform, such as public support that can be translated into more votes in the next elections (Rahat, 2006; 2008), what Renwick (2009: 286) calls the act-contingent vote-seeking behavior of politicians. But even lesser gestures by politicians can serve the cause of reform. Pressure may lead politicians who oppose reform to establish a committee, or even conduct a referendum, with the intention of showing responsiveness to the public without letting reform be adopted (Massicotte, 2008). However, reformers may 4 As examples see: Boix (1999) and the critique by Andrews and Jackman (2005) and by Blais, Dobrzynska and Indridason (2004); Bueno de Mesquita (2000) and the critique by Rahat (2004; 2006); Benoit (2004) and the critique by Van der Kolk (2007). See also Katz (2007).

29 28 sometimes skillfully use these cracks in the vested interests bastion: an expert committee might turn into a legitimizing vehicle for reform rather than the usual expected cemetery; a referendum might become an anti-hegemonic tool rather than a tool to legitimize the status quo (e.g. New Zealand, and almost also in British Columbia). An additional tactic that can help to calm some vested interests is to offer a delay in the implementation of reform (Rahat, 2008; Renwick, 2009). This step allows those in power not only to enjoy the system one more time and still be seen as reformers who are ready to give up on their vested interests for the public interest, but is also gives them time to adapt to the new rules from their advantageous position within the political system. It should be remembered that vested interests might be diverse, and even contradicting, when looking at different levels (personal, factional, partisan, national), at various time horizons (short vs. long term), and with contrary assumptions about future voting behavior (voter loyalty or volatility). Reformers can, therefore, manipulate information, making those interests that are served by the reform more prominent than those that are likely to be hurt. Another way to deal with the challenge of vested interests is to inject uncertainty concerning the future results of reform, even up to the point where it will be unclear who will win and who will lose. When it is clear who wins and who loses from electoral reform, it is hard to convince the needed majority to support it, especially